Getting Started
Discovery & Architecture
How SPECTRALE discovery identifies components, relationships, and external boundaries from available project evidence.
Understand what your software is made of.
Discovery is the foundation of SPECTRALE.
It identifies technical components and relationships from available project evidence, helping you build a clearer picture of your software environment.
What can Discovery identify?
Depending on the connected discovery source and available evidence, SPECTRALE may identify:
Applications and services
Backend applications, frontend applications, workers, and other software services.
Data systems
Databases, storage systems, caches, and related infrastructure.
Interfaces and dependencies
APIs, service connections, package dependencies, and communication relationships.
Development infrastructure
Repositories, build configurations, deployment references, and related technical resources.
External components
Third-party services and external systems referenced by the architecture.
The architecture graph
SPECTRALE represents architecture through two fundamental concepts.
Nodes
Nodes represent architectural components.
Examples include:
- Application Service
- PostgreSQL Database
- API Gateway
- GitHub Repository
- External API
Edges
Edges represent relationships between components.
Examples include:
- A service depends on a database.
- An application communicates with an API.
- A repository contains an application.
- A service references an external system.
Together, nodes and edges create a navigable model of the software architecture.
Discovered external components
SPECTRALE can identify external services referenced within your architecture.
These components appear in the project's discovery results and can help reveal dependencies beyond your immediate codebase.
External component discovery is informational.
It does not automatically establish authenticated connections, retrieve private third-party data, or grant access to external accounts.
Working with multiple projects
SPECTRALE organizes architectural information into projects.
Each project has its own discovery context and architectural representation.
This separation also matters for MCP: an AI agent should only receive information from projects it has been explicitly authorized to access.